Veterinary Microbiology 136 (2009) 326334

Contents lists available at ScienceDirect

Veterinary Microbiology
journal homepage: www.elsevier.com/locate/vetmic

Characterisation of an extracellular vibriolysin of the sh pathogen

Moritella viscosa
Bryndis Bjornsdottir a, Olafur H. Fridjonsson b, Steinunn Magnusdottir b,
Valgerdur Andresdottir a, Gudmundur O. Hreggvidsson b,c, Bjarnheidur K. Gudmundsdottir a,*
a
b
c

Institute for Experimental Pathology, University of Iceland, Keldur, Vesturlandsvegi, 112 Reykjavik, Iceland
Matis-Prokaria, Gylfaot 5, 112 Reykjavik, Iceland
Department of Biology, University of Iceland, Sturlugotu 7, 101 Reykjavik, Iceland

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 29 September 2008
Received in revised form 25 November 2008
Accepted 28 November 2008

Moritella viscosa causes winter ulcer disease in salmonids. The aim of the present work was
to isolate and partially characterise an extracellular peptidase from M. viscosa, and to study
its role in virulence. The peptidase, termed MvP1, was a 38-kDa metallopeptidase
produced in late exponential growth. The optimum temperature for MvP1 was 40 8C, but
the enzyme was active over a broad range of temperatures. MvP1 was non-lethal to
salmon at concentrations up to 0.22 mg/g sh, but extracellular products were lethal to
salmon. MvP1 degraded casein, gelatin and collagen from lumpsh skin. It caused
considerable tissue necrosis and hemorrhages at the site of injection, and affected cellcell
adhesions in EPC and BF-2 cell lines, but was not highly cytotoxic. The peptidase partially
degraded sh IgM heavy chain but was non-hemolytic. The mvp1 gene was sequenced and
encoded a 734-aa polypeptide containing a signal sequence, an N-terminal propeptide, a
mature peptidase domain and a C-terminal propeptide. The MvP1 propeptide undergoes
both N-terminal and C-terminal processing and different C-terminal processing results in
the formation of several active isoforms of the mature peptidase. The catalytic domain
showed highest sequence similarity with several vibriolysins (EC 3.4.24.25) originating
from Pseudoalteromonas strains, showing up to 80% aa identity. The results indicate that
MvP1 is a previously unknown vibriolysin that might affect M. viscosa virulence by aiding
in the invasion and dissemination of the bacterium in its host, by causing tissue
destruction.
2008 Elsevier B.V. All rights reserved.

Keywords:
Moritella viscosa
MvP1
Extracellular products
Vibriolysin
Virulence factor

1. Introduction
Moritella viscosa causes winter ulcer disease in sh
reared at low temperatures in North Atlantic countries,
mainly in Norway (reviewed in Gudmundsdottir and
Bjornsdottir, 2007). Large ulcers often appear on infected
sh, and hemorrhages and tissue necrosis are detectable
internally (Lunder et al., 1995; Bruno et al., 1998). The

* Corresponding author. Tel.: +354 5855100; fax: +354 5674714.

E-mail address: bjarngud@hi.is (B.K. Gudmundsdottir).
0378-1135/$ see front matter 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.vetmic.2008.11.020

disease causes signicant mortalities and nancial losses

in Atlantic salmon, Salmo salar (L.) and rainbow trout,
Oncorhynchus mykiss (Walbaum), and causes infections in
Atlantic cod, Gadus morhua (L.). Furthermore, the bacterium has been isolated from several other sh species
(reviewed in Gudmundsdottir and Bjornsdottir, 2007).
Lipooligosaccharides and a 1719 kDa outer membrane
antigen have recently been shown to be protective
antigens (Heidarsdottir et al., 2008).
Pathogenic bacteria produce a variety of virulence
factors that affect their hosts, allowing the bacteria to
multiply and produce a disease. Several extracellular

B. Bjornsdottir et al. / Veterinary Microbiology 136 (2009) 326334

bacterial peptidases, mainly metallopeptidases, have been

shown to have either direct or indirect roles in pathogenesis and virulence, although most are known for nutrient
acquisition for bacterial growth (Miyoshi and Shinoda,
2000). The Pseudomonas aeruginosa elastase and Vibrio
vulnicus and Vibrio cholerae vibriolysins have been
implicated with bacterial virulence. They cause various
actions, such as tissue destruction and hemorrhages, and
are believed to play roles in colonization, invasiveness, and
dissemination within the host (Booth et al., 1984;
Finkelstein et al., 1992; Komori et al., 2001; Silva et al.,
2003; Miyoshi, 2006).
This paper describes the isolation and characterisation
of a previously unknown extracellular peptidase from the
sh pathogenic bacterium M. viscosa.
2. Materials and methods
2.1. Bacterial strain, growth conditions, and production of
ECP
Moritella viscosa, strain K58, was isolated and identied
in 1992 at the Institute for Experimental Pathology,
University of Iceland, from the head kidney of diseased
Atlantic salmon from a farm in SW-Iceland (Benediktsdottir et al., 1998, 2000). The strain was passaged in
salmon prior to the study, by reisolation from kidney of i.p.
infected sh. The bacterium was grown on 5% horse blood
agar supplemented with 2% NaCl at 4 8C for 4 days. M.
viscosa growth curve was determined by averaging three
identical batch cultures cultivated in Brain-Heart infusion
broth (BHI, BD) containing 1.5% NaCl (BHI-NaCl) at 9 8C,
with agitation (200 rpm). Growth was estimated in log
CFU/ml by serial dilution and plate counting.
Extracellular products (ECP) were isolated from batch
cultures grown at 4 or 9 8C for 5 days by centrifugation, and
then ltered (Whatman, 22 mm). For isolation of ECP, used
for toxicity testing, M. viscosa was cultivated using the
cellophane overlay method (Liu, 1957) on Tryptone soya
broth (Difco) agar with 1.5% NaCl at 15 8C for 3 days.

327

Remaining caseinolytic activity was determined using

azocasein assay.
2.3. Peptidase purication
ECP proteins were precipitated by adding ammonium
sulphate to 85% saturation. The precipitate was recovered
by centrifugation and dissolved in 20 mM Tris (pH 8).
Precipitated proteins were desalted using PD-10 columns
(Amersham Biosciences), equilibrated with the same
buffer at 4 8C. Desalted samples were loaded onto a
MonoQ HR 5/5 column (Pharmacia), preequilibrated with
20 mM Tris (pH 8). The column was eluted at 15 8C with a
linear gradient of 00.8 M NaCl in the same buffer. Eluate
fractions were collected and analyzed for caseinolytic
activity. Pooled fractions containing the highest caseinolytic activities were concentrated by ultraltration
using Amicon Ultra-15 lter (Millipore, MWCO 10,000) at
4 8C. Concentrated fractions were further fractionated on a
Superdex 200 HR 10/30 column (Amersham Pharmacia
Biotech), equilibrated and eluted with 50 mM Na-phosphate buffer (pH 7), containing 15 mM NaCl (SD-200
buffer) at 15 8C. Fractions with major caseinolytic activities
were pooled and used as the puried enzyme, following
analysis with SDS-PAGE and Western blotting.
2.4. Electrophoresis and isoelectric focusing
Standard SDS-PAGE was carried out according to the
Laemmli method (Laemmli, 1970) in 12% polyacrylamide
gels and protein bands visualised by silver staining (Silver
Stain Plus Kit, BioRad). Zymography was performed as
previously described (Gudmundsdottir, 1996), using 0.03%
casein, 0.1% collagen (isolated from lumpsh skin), or 0.1%
gelatin as substrates. Gels were incubated for 2 h at 30 8C.
Isoelectric focusing was carried out in Phast system
(Pharmacia), using PhastGel IEF 46.5 (GE healthcare), and
standards pH 310 (Pharmacia). Gels were xed in 20%
trichloroacetic acid and visualised with silver staining.
2.5. Immunoreactivity of MvP1 isoforms

2.2. Proteolytic activity, protein content, and inhibition

studies
Caseinolytic activity was assayed using azocasein
(Sigma) as previously described (Gudmundsdottir, 1996),
and incubated at 30 8C for 60 min. Samples were assayed in
duplicate and each duplicate measured twice. One unit (U)
of caseinase activity was dened as an increase in
absorbance of 0.01 under the assay conditions.
Protein content was measured with a protein assay kit
(Coomassie Plus, Pierce) and bovine serum albumin
(Sigma) used for plotting a standard curve. Absorbance
was measured at 590 nm and all samples measured in
duplicate.
The pure peptidase was incubated with inhibitors for
10 min at 22 8C. Stock solutions of the dissolved metallopeptidase inhibitors EDTA (Merck) and OPA (1,10-phenanthroline) (SigmaAldrich) and the serine peptidase
inhibitor PMSF (Sigma) were prepared and added to the
MvP1 peptidase in nal concentrations of 1 or 10 mM.

Isolated MvP1 was electrophorised and visualised using

Western blotting as described previously (Gudmundsdottir et al., 2003). Blotted membranes were incubated with
polyclonal mouse antibodies against the 38 kDa MvP1
isotype (I1, see denition in Section 3.2) that were
prepared in mouse ascitic uid using a previously
described method (Gudmundsdottir et al., 2003), and with
an estimated dose of 10 mg MvP1 injected each time.
Antibodies were diluted 1:2000. Washed membranes were
then incubated with goat anti-mouse antibodies conjugated with alkaline phosphatase (Dako), diluted 1:1000.
2.6. Optimal temperature and thermostability
For estimating the optimal temperature for MvP1, the
catalytic activity of the pure peptidase was measured at
different temperatures between 5 and 80 8C for 1 h, by the
azocasein assay. For thermostability testing the puried
protein was incubated at 30, 40, 50 and 60 8C for 30 min in

328

B. Bjornsdottir et al. / Veterinary Microbiology 136 (2009) 326334

SD-200 buffer and then centrifuged before performing the

azocasein assay. Samples were measured in duplicates.
2.7. Toxicity testing
Healthy unvaccinated Atlantic salmons from a commercial sh farm were used for toxicity testing of M.
viscosa ECP, and isolated MvP1. The sh were 46 g (6 g)
and kept in fresh water in a 0.15 m3 tank at 8 8C.
Intramuscular (i.m.) and intraperitoneal (i.p.) injections
with 0.1 ml ECP (2 mg/g sh (70 U)) or 0.2 ml MvP1 (up
to 0.22 mg/g sh (65 U)) were performed. Two to six
sh were included in each group and groups were
identied with Alcian blue markings. The sh were
monitored for up to 9 days and complete necropsy
performed on all sh. Fish were injected with SD-200
buffer as control.
2.8. Cytotoxicity assay
Cytotoxic activity of M. viscosa ECP and puried MvP1
was determined using epithelioma papulosa carpio (EPC)
and bluegill fry broblast (BF-2) cells. Conuent cell
monolayers were prepared in 96-well at-bottom cell
culture plates containing 150 ml Eagles minimal essential
medium (Eagles MEM, Gibco), with 10% foetal bovine
serum (Gibco), penicillin (100 IU/ml), and streptomycin
(100 mg/ml), and cultured at 23 8C for 20 h. Then, medium
was removed from wells and 150 ml of fresh Eagles MEM
with or without 10% foetal bovine serum added, as well as
40 ml of different dilutions of ECP (22175 mg/ml) or MvP1
(330 mg/ml), containing equal caseinolytic activities
(50 U in highest concentrations). Each treatment was
carried out in triplicate and SD-200 buffer used as control.
The cells were incubated for further 24 h at 15 8C and
microscopically examined.
2.9. Proteolytic digestion of IgM
Isolated salmon and cod IgM (4 mg) (Magnadottir,
1998) were incubated in duplicate for 24 h at 9 8C with M.
viscosa ECP (1.2 mg) or isolated MvP1 (0.12 mg), each
containing caseinolytic activity of 65 U. Incubation with
SD-200 buffer was used as control. Samples were then
electrophorised using standard SDS-PAGE and IgM visualised using Western blot analysis as described in Section
2.5, with the exception that membranes were incubated
with polyclonal mouse antibodies against salmon or cod
IgM (Magnadottir, 1998), diluted 1:2000.
2.10. Hemolytic assay
Hemolytic activity was determined using salmon and
sheep erythrocytes. Serial dilutions of M. viscosa ECP and
isolated MvP1 were incubated for 5 and 24 h at 22 8C, in
0.5% washed red blood cells suspended in phosphate
buffered saline (PBS, Sigma). Each dilution was carried out
in duplicate. PBS was used to calculate 0% lysis and dH2O to
calculate 100% lysis. Absorbance of the supernatant was
measured at 405 nm, and hemolysis of each sample
dilution calculated.

2.11. N-terminal sequence analysis

MvP1 was blotted on to a polyvinylidene uoride
membrane (Millipore) and the N-terminal aa sequence of
the 38 kDa isotype (I1, see denition in Section 3.2)
determined at the Institut Pasteur, Paris by automated
Edman degradation.
2.12. DNA sequencing and computer analysis
Degenerate primers were designed for PCR amplication of a sequence from mvp1, the gene that encodes
MvP1. Construction of primers was based on N-terminal
sequence analysis and the third zinc ligand motif of
thermolysin peptidases. A list of primers used in this
study is given in Table 1. The 508-nt PCR product was
cloned into pCR12.1-TOPO1 (Invitrogen) and sequenced
by the dideoxy method using BigDye1 Terminator
sequencing kit (Applied Biosystems) and universal
M13 primers. Then, a series of nested PCR amplications
and subsequent sequencing was performed to determine
anking sequences. Gene-specic 50 biotin-labelled
primers and arbitrary primers (Arb1 or Arb2) were used
for rst round PCR amplication. PCR products were
puried using the Qiaquick PCR purication Kit (Qiagen)
and biotin labelled PCR products further puried using
Dynabeads1 M-280 Streptavidin (Dynal Biotech). Second round PCR amplication was performed using
nested gene-specic primers, which were designed to
bind upstream of the previous gene-specic primers, and
also a nested primer to Arb1 and Arb2 (Arb3). PCR
products were puried using GFXTM PCR, DNA and Gel
Band Purication Kit (Amersham Bioscienses), cloned
into pCR14-TOPO1 (Invitrogen) and transformed into
One Shot1 TOP10 competent Escherichia coli (Invitrogen). Cloned inserts were PCR amplied using M13
primers and cycle sequenced. The sequence information
was used to generate new gene-specic primers for the
next nested PCR round until the complete gene and
anking regions were obtained. All PCR cycling conditions followed standard laboratory practices. For conrmation of mvp1 sequencing, the gene and its anking
regions were PCR amplied from genomic DNA, using
primers containing EcoRI and BamHI recognition
sequences, and cloned into pUC19 (Invitrogen), using
EcoRI and BamHI. Chemically competent DH5a E. coli
cells (Invitrogen) were transformed with the resulting
ligation product, and both strands of the cloned insert
sequenced. DNA sequences were analyzed using
Sequencher 4.6 (Scientic & Educational Software).
Nucleotide acid and translated amino acid sequences
of the mvp1 gene and anking sequences have been
deposited in the GenBank database under accession
number EU876833. The NCBI Basic Local Alignment
Search Tool (BLAST) was used for comparing protein
sequences and CD search for detecting conserved
domains. Putative open reading frames (ORF) were
identied by the SoftBerry FGENESB program. The
ProtParam program (ExPASy) was used for MW computation and SignalP 3.0 (CBS) for predicting the location of
a signal peptide cleavage site.

B. Bjornsdottir et al. / Veterinary Microbiology 136 (2009) 326334

329

Table 1
Oligonucleotide primers used in the study.
No.

Name

forward (f)
reverse (r)

50 - 30

Application

1
2
3

mvp1-N5-f
mvp1-I492-r
Arb1

f
r
f/r

GGMTTYGGTGGTAACGARAAAACRGG
GCTTCACCWGCCATATCAGARAAYGCTTC
GGCCACGCGTCGACTAGTACNNNNNNNNNNGATAT

Amplication of a 508-nt fragment of mvp1

4
5
6

Arb2
Arb3
mvp1-1bio

f/r
f/r
r

GGCCACGCGTCGACTAGTANNNNNNNNNNACGCC
GGCCACGCGTCGACTAGTAC
ATTACCAGACTGCCCTACATTC

7
8

mvp1-2bio
mvp1-3

f
r

TTCAGCTCATGAAGTTAGTC
CAAGTATCCGAAATCAGTACC

9
10
11
12
13
14
15
16
17
18
19

mvp1-4
mvp1-5bio
mvp1-6bio
mvp1-7
mvp1-8
mvp1-9bio
mvp1-10bio
mvp1-11
mvp1-12
mvp1-13bio
mvp1-15a

f
r
f
r
f
r
f
r
f
r
r

GGTTTATGAGGCTAAATCAG
CAACAGCACCATGAGCATTC
GTACTGACAGTGGTGATGAG
CCGTCAGAATAACGAGATATAATT
CAGGTGCCAGTGACATTC
ACACCACCATCACTATCG
CGGTTACAGTACTGCATCTG
GACCACACTAAGACCTTC
TGATGTGGAGCTAAATATTCA
GTTGTTGCTGAAGATATTGC
GCTGATAATTGTTGAATATTGTC

20
21

mvp1-EcoRI
mvp1-BamHI

f
r

CGGAATTCGACGTACCGCACCAAATCC
CGGGATCCTATTAGCTAAATTATCTGCTG

Ligation of mvp1 and anking regions into pUC19

Restriction sites for cloning are underlined

22
23
24
25
26
27
28
29

mvp1-425
mvp1-990
mvp1-1478
mvp1-3100
mvp1-2654
mvp1-2190
mvp1-1794
mvp1-446

f
f
f
r
r
r
r
r

CAGTCGCTTACAATTTTCAGTT
TCTCGATAGTGATGGTGGTG
TGAATGTAGGGCAGTCTGG
GCCAACGTCGATAGAGAAAC
GTATAGCTGCCGACCAAACT
GCTCCAATAGAGTTGGTTGG
GTACGATTCACCATCACCAA
AACTGAAAATTGTAAGCGACTG

Sequencing of mvp1 and anking regions in pUC19.

Primers no. 8, 9, 17 and 19 were also used

2.13. Licenses
The study was approved by and performed according to
the Icelandic Animal Research Authority (approval no. YDL
03080041/023BE).

Arbitrary primers for outward sequencing

of mvp1 and anking regions
Nested primer for Arb1 and Arb2
Specic primers for outward sequencing of
mvp1 and anking regions.
Even numbered primers are designed for
upstream amplication, odd numbered
primers are designed for downstream
amplication. Primers marked bio are 50
biotin labelled and other primers are nested

and 15 8C, where the same band pattern was observed.

Polyclonal antibodies raised against the 38 kDa MvP1 band
bound to all caseinolytic bands detected in ECP (data not
shown).
3.2. MvP1 isolation

3. Results
3.1. M. viscosa cultivation and peptidase production
M. viscosa generation time in BHI-NaCl at 9 8C was 2 h
and 48 m, as calculated from plate counts in the log phase
(11 and 17 h timepoints) of all three cultures (Fig. 1). Under
these conditions the cultures reached a maximum of
1.2  1010 CFU/ml after 96 h of cultivation.
Peptidase production in the ECP was observed at the
end of the exponential growth phase. The activity was
detected after 35 h of cultivation by casein zymogram and
after 43 h when using azocasein assay. Maximum caseinase activity (71 U) was measured at the nal timepoint at
111 h and two caseinolytic bands were detected in a
zymogram of ECP collected at 35 h or later. Occasionally, a
third faint band of a smaller size was also identied in
stationary phase ECP (data not shown). Peptidase production was also identied in ECP from cultures grown at 4

Following ECP precipitation, several caseinolytic bands of

varying numbers ranging from 24 to 40 kDa were detectable
in a zymogram (data not shown). After ion-exchange FPLC,

Fig. 1. Growth curve of M. viscosa K58 (*) ( S.D.), and production of

caseinolytic peptidase (*) in BHI-NaCl at 9 8C. The gure shows the average
from three parallel batch cultures.

B. Bjornsdottir et al. / Veterinary Microbiology 136 (2009) 326334

330

Fig. 2. Purication of the extracellular M. viscosa K58 MvP1 peptidase. (A) MonoQ HR 5/5 anion exchange chromatography of precipitated ECP proteins. (B)
Superdex 200 gel ltration chromatography of the major caseinolytic peak (fractions 1415) from A. (C) Three active MvP1 isotypes (fractions 3032) from
B, termed I1, I2 and I3. The I1 and I2 isotypes appeared as 38 and 36 kDa bands in SDS-PAGE, but as 40 and 38 kDa bands in casein zymogram, respectively.
Both methods produced a 35 kDa band of I3. Lane S, molecular weight standards; lane 1, silver staining; lane 2, casein zymogram.

Table 2
Summary of the purication of M. viscosa MvP1 peptidase.
Purication step

Total protein (mg)

Total proteolytic units

Specic activity (units/mg)

Recovery (%)

Relative purication

Extracellular products
1 Amm. sulf. precipitation
2 PD-10 gel ltration
3 Mono Q ion exchange
4 Ultraltration
5 Superdex 200 gel ltration

17.68
2.84
2.14
0.46
0.36
0.05

3224
473
464
219
66
51

4
54
50
221
364
2020

100
13.0
12.8
6.0
1.8
1.4

1
14
13
56
92
510

ECP proteins eluted as several peaks with caseinolytic

activity eluting in two peaks (Fig. 2A). The major caseinolytic
peak (fractions 1415) typically produced three protein
bands, hereafter termed MvP1 isotypes I1, I2 and I3. The I1
and I2 isotypes appeared as 38 and 36 kDa bands in SDSPAGE, but as 40 and 38 kDa bands in casein zymogram,
respectively. Both methods produced a 35-kDa band of I3
(Fig. 2C). The minor peak (fraction 20) produced mainly the
smaller caseinolytic bands. After further separation of the
major peak by gel ltration on Superdex 200, the MvP1
peptidase eluted in one peak (fractions 3031) (Fig. 2B).
Silver staining of the peak revealed one to three protein
bands (I1, I2 and/or I3) varying between purications. The
protein bands were all conrmed as MvP1 by Western
blotting and were all active (Fig. 2C). Specic activity of the
three isolated MvP1 isotypes was calculated 1020 U/mg
isolated protein, using caseinase units. Total recovery of
peptidase activity from ECP was 1.4% and the relative
purication was 510-fold (Table 2). The percentage of MvP1
in total extracellular proteins was estimated to be from 10 to
17%, based on the specic activity of MvP1.
3.3. MvP1 characterisation
The MvP1 peptidase degraded casein, lumpsh skin
collagen and gelatin, and the same band pattern was

observed in zymograms containing the three different

substrates. The metallopeptidase inhibitors OPA and EDTA,
at 1 mM concentrations, reduced activity by 73 and 16%,
respectively. At 10 mM, OPA inhibited MvP1 activity
completely but 21% of the activity still remained after
treatment with 10 mM EDTA. The serine peptidase inhibitor
PMSF did not affect MvP1 activity at 1 mM concentration.
Optimum reaction temperature for the puried MvP1
peptidase was around 40 8C (Fig. 3). The peptidase showed

Fig. 3. Temperature effect on MvP1 activity. (*) Activity range and

optimum temperature measured by incubating isolated MvP1 in
azocasein sample mixture at a given temperature for 60 min. (*)
Thermostability measured by incubating isolated MvP1 at a given
temperature for 30 min prior to measuring azocasein activity. Activity at
40 8C was set as 100%.

B. Bjornsdottir et al. / Veterinary Microbiology 136 (2009) 326334

activity between 5 and 60 8C and at 5 8C 35% of the relative

activity still remained. At temperatures of 45 8C or higher,
the activity declined rapidly. MvP1 thermostability is also
shown in Fig. 3. After incubation at 50 8C for 30 min,
peptidase activity dropped signicantly, and at 60 8C no
residual activity was measured. In a sample of the puried
MvP1 peptidase, containing the I1 and I2 isotypes, two
isoelectric points could be detected with calculated pI of
4.4 and 4.2.
Isolated MvP1 was non-lethal to salmon at concentrations up to 0.22 mg/g sh. However, the peptidase
caused considerable hemorrhages and extensive tissue
necrosis at the site of injection in concentrations as low
as 0.06 mg/g sh, and ascites formation was detected
after i.p. injection. The ECP of M. viscosa were lethal to
salmon, at 2 mg/g sh, causing death in all injected sh
within 72 h. Extensive tissue necrosis and hemorrhages
were seen around the site of injection, and hemorrhages
were also seen in various internal organs and tissues, in
both i.m. and i.p. injected sh. Ascites was seen in the
peritoneal cavity of all sh, and petechial hemorrhages
in liver and perivisceral fat were seen in two out of four
sh.
Cytotoxic activity of M. viscosa ECP and puried MvP1
against EPC and BF-2 cells was tested. Each dilution of
ECP and isolated MvP1 contained equal caseinolytic
activities. Serum decreased the toxic effects of ECP and
MvP1, but did not hinder them completely. Effects seen
in both cell types were similar. ECP caused cell death,
ruptures in the cell monolayer, and detachment from the
edges of the wells. The MvP1 peptidase did not cause
prominent cell death, which was only detectable at the
higher concentrations. However, MvP1 caused detachments from well edges and ruptures in the cell
monolayer.
Isolated MvP1 partially digested both salmon and cod
IgM heavy chain, but IgM was not affected when incubated
with ECP, containing equal caseinolytic activity (56 U)
(Fig. 4).
MvP1 was non-hemolytic against salmon and sheep
erythrocytes, but ECP were strongly hemolytic against

331

salmon erythrocytes (1:126 dilution caused 20 and 89%

lysis after 5 and 24 h, respectively), and weakly hemolytic
against sheep erythrocytes (1:4 dilution caused 2 and 18%
lysis after 5 and 24 h, respectively).
3.4. mvp1 sequencing
In total, 3.562-nt were sequenced. A potential mvp1 ORF
of 2.205-nt encoding a 734-aa polypeptide with calculated
MW of 79.095 was found (Fig. 5). A putative signal
sequence was detected, with a predicted cleavage site after
Ala-25. The previously determined aa sequence for the Nterminal end of MvP1 (ADATGFGGNEKTGKYHYGTDF) was
found within the polypeptide, starting at Ala-211. The
highly conserved HEXXH zinc-binding motif of metallopeptidases was found in mvp1 (aa 351355), and also the
third zinc ligand motif GXXNEXXSD (aa 371379). A
possible ShineDalgarno (SD) sequence (ggagac) was
detected upstream of the predicted Met start codon (nt
10 to 5). Another, less possible, Met start codon was
also found (nt 12 to 10), that partially overlapped with
the predicted ribosomal binding site. An inverted repeat
sequence, characteristic of a Rho-independent transcription determinator was found downstream of the stop
codon (nt 22272243).
Several conserved domains were identied within the
MvP1 prepropeptide, which could be divided into four
regions; a signal peptide, an N-terminal propeptide that
contains both an FTP domain (aa 64114) and a PepSy
domain (aa 127205), the mature peptidase, consisting of
an M4 domain (aa 217360) and an M4 C-terminal domain
(aa 362506), and nally a C-terminal propeptide, containing two PPC domains (aa 543612 and 650719).
The calculated size of the protein sequence from the Nterminal end was 56 kDa, whereas the puried peptidase
was seen as a 38 kDa band or smaller by SDS-PAGE. This
suggests that MvP1 undergoes polypeptide removal from
the carboxy terminal by at least 18 kDa. BLAST search
revealed that the translated protein sequence of the
predicted mature peptidase unit (aa 217506) had highest
similarity with vibriolysin (EC 3.4.24.25). It showed

Fig. 4. Western blot analysis of salmon and cod IgM. Lane 1, untreated IgM; lane 2, IgM incubated with MvP1 peptidase; lane 3, IgM incubated with ECP; lane
S, molecular weight standards as indicated in kDa. Incubation was at 9 8C for 24 h.

332

B. Bjornsdottir et al. / Veterinary Microbiology 136 (2009) 326334

Fig. 5. Nucleotide sequence of the mvp1 gene. Deduced amino acids of the MvP1 prepropeptide are shown below the rst nucleotide of each codon. The rst
nt of the Met initiation codon is set as +1 and the initiation Met is set as the rst aa. A putative SD sequence is double-underlined and the predicted signal
peptide cleavage site is shown with a vertical arrow. The N-terminal amino acid sequence of MvP1 is underlined. The HEXXH zinc-binding motif is shaded
and the third zinc ligand-motif is boxed. An inverted repeat sequence, characteristic of a Rho-independent transcription determinator, is indicated by
convergent arrows.

highest percent sequence similarity with the MprI

peptidase of Pseudoalteromonas piscicida (80% sequence
identity, accession number BAB79615) and ve other
Pseudoalteromonas vibriolysins (7779% sequence identity,
accession numbers ACI28452, BAG70540, BAB85124,
ABL06977, and EAR27939). Signicant sequence similarity
of the complete mvp1 ORF extended throughout the entire
ORFs of the Pseudoalteromonas peptidases. Two partial
putative ORFs were also detected in the 3.562-nt sequence.
A 483-nt sequence was found 59-nt downstream of mvp1,
and showed highest aa similarity (40% sequence identity)
with an aminopeptidase of P. piscicida (MprII, accession
number BAB79616). A signal peptide with a predicted
cleavage site after Ala-20 was detected in the sequence. A
212-nt sequence found 603-nt upstream of mvp1, encoded
on the opposite strand and transcribed in the other
direction, was identical with a predicted proline amino-

peptidase sequence from Moritella sp. PE36 (accession

number EDM65361).
4. Discussion
This paper describes the isolation and partial characterisation of MvP1, an extracellular metallopeptidase
from the sh pathogen Moritella viscosa. The peptidase has
virulence related activities and is able to degrade host
tissue and IgM. It shows highest aa sequence similarity
with Pseudoalteromonas vibriolysins. We also showed that
the M. viscosa ECP are lethal to salmon.
MvP1 was the only detectable caseinolytic enzyme
secreted by M. viscosa at 4, 9 and 15 8C. Several active
isoforms of MvP1 were identied, mainly appearing after
precipitation and other handling, indicating peptidase
degradation of a single polypeptide. The isotypes also

B. Bjornsdottir et al. / Veterinary Microbiology 136 (2009) 326334

appeared through natural processing during storage of

ECP. The isolation of two or more isotypes of the same
vibriolysin has been described previously (Wu et al., 1996;
Miyoshi et al., 2002).
Total recovery of peptidase activity after isolation was
1.4%, and most of the activity was lost during the
precipitation step, where only 13% of the activity measured
in freshly collected ECP was recovered. The MvP1 is an
abundant protein in the ECP of M. viscosa, as the estimated
percentage of the peptidase ranged from 10 to 17%. To our
knowledge, this is the rst protein isolated and characterised from M. viscosa.
Optimum reaction temperature for isolated MvP1 was
around 40 8C, but activity was detected over a broad
temperature range. The relatively high activity of MvP1 at
low temperatures probably reects the low temperatures
at which the bacterium is found and causes infections. The
peptidase was not highly thermostable, showing
decreased stability at 50 8C and no residual activity at
60 8C.
MvP1 was non-lethal to salmon, but caused extensive
tissue necrosis and hemorrhages. However, ECP containing equal caseinolytic activities, were lethal to salmon
and produced internal disease signs similar to those seen
in sh infected with live bacteria. The ECP were isolated
from a cellophane overlay culture, representing partially
restricted nutrition, which resembles the host environment. The LD50 of the ECP of a Norwegian M. viscosa
strain, F288/95, has been estimated in salmon, 0.3 mg/g
sh, (Bjornsdottir, B., et al., unpublished results).
Thus, MvP1 does not appear to be the lethal factor in
M. viscosa ECP, or its effects may be enhanced by other
unknown factors in the ECP. The necrotic and hemorrhagic damage caused by MvP1 may be due to the
digestion of structural components, and possibly affect
the development of ulcers that are often associated
with M. viscosa infections. It has been proposed that
bacterial peptidases, causing tissue damage and degradation of host tissues, play roles in pathogenesis by
facilitating invasion of the pathogen (Miyoshi and
Shinoda, 2000).
The isolated MvP1 peptidase did not cause prominent
cell death in the two sh cell lines tested, EPC and BF-2, but
seems to affect cellcell adhesions. Disruption of epithelial
cell adhesions can potentially aid the invasion and
dispersion of M. viscosa in its host. M. viscosa ECP were
more cytotoxic than isolated MvP1, and caused cell death
and prominent morphological changes, indicating that
MvP1 is not the major cytotoxic factor.
Salmon and cod IgM heavy chain was partially
degraded by isolated MvP1, which might have an effect
on the host immune system. The MvP1 peptidase may
thus contribute to the pathogenesis of M. viscosa by
obstructing the hosts immune response. However, ECP
containing equal caseinolytic activity did not degrade
IgM. Thus, it seems that one or more factors in the ECP
hindered the degradation. Whether this effect would be
observed in vivo is not known. MvP1 was non-hemolytic
against salmon and sheep erythrocytes, while hemolytic
activity was detected in the ECP, especially against salmon
erythrocytes.

333

The structural gene of the MvP1 peptidase was

sequenced. Sequence alignments indicate that MvP1 is a
member of the thermolysin family of zinc metallopeptidases. Inhibition studies conrmed the metallopeptidase
nature of MvP1. The sequence contained the highly
conserved HEXXH zinc-binding motif of metallopeptidases
(Rawlings and Barrett, 1995), and the third zinc ligand
motif GXXNEXXSD, that is characteristic of the thermolysin family (Miyoshi and Shinoda, 2000). The translated aa
sequence showed highest similarity to vibriolysin peptidases (EC 3.4.24.25) that belong to the thermolysin family.
Highest similarity was found to several vibriolysins from
different Pseudoalteromonas species. According to the
MEROPS database, vibriolysins originating from several
non-pathogenic environmental bacteria, and pathogenic
vibrios, causing diseases in both humans and marine
organisms, have been described (Rawlings et al., 2008).
Several vibriolysins have been implicated with virulence,
although their functions are not fully understood, and they
do not seem to be essential for bacterial virulence
(Finkelstein et al., 1992; Miyoshi et al., 2002; Silva et al.,
2003; Denkin and Nelson, 2004; Miyoshi, 2006).
The MvP1 prepropeptide can be divided into four
regions, a signal peptide, an N-terminal propeptide, a
mature peptidase, and a C-terminal propeptide. The
determined N-terminal sequence of MvP1 started at aa
211, showing that an N-terminal propeptide of about
20 kDa is cleaved off to form the mature peptidase. Further
N-terminal processing is unlikely, as the N-terminal end is
found to be similar to previously described N-terminal
ends of several metallopeptidases, and because it is located
only 6 aa upstream of the predicted mature peptidase
motif. Also, it seems that a C-terminal propeptide of at least
18 kDa is cleaved off the mature peptidase, since the
apparent size of isolated MvP1 is 38 kDa or smaller,
whereas the calculated size of the propeptide starting from
the N-terminal end is 56 kDa. The presence of various
active MvP1 isoforms of different sizes suggests that the Cterminal propeptide is cleaved off at different sites and
results in the formation of the different isoforms of the
mature peptidase. Propeptide processing at both the Nand C-terminal end has been described for several
vibriolysins (David et al., 1992; Miyoshi et al., 1997; Lee
et al., 2002; Miyoshi et al., 2002).
The N-terminal propeptides of thermolysins are known
to inhibit peptidase activity in the cytoplasm and assist in
protein folding (Tang et al., 2003; Chang et al., 2007). Two
conserved domains were found within the N-terminal
propeptide of MvP1. A FTP domain (Fungalysin/Thermolysin propeptide motif), which is believed to inhibit the
peptidase and/or have chaperone activity, and a PepSY
domain (Peptidase propeptide and YPEB domain), which is
proposed to have an inhibitory function (Yeats et al., 2004;
Marchler-Bauer et al., 2007; Demidyuk et al., 2008). The
mature peptidase region consists of a M4 domain and a M4
C-terminal domain that is unique for the thermolysin
family (Rawlings et al., 2008). Two PPC domains (bacterial
pre-peptidase C-terminal domains) were detected in the Cterminal propeptide of MvP1, which are often not found in
mature peptidases. The function of the C-terminal
propeptide of thermolysin like peptidases is not known,

334

B. Bjornsdottir et al. / Veterinary Microbiology 136 (2009) 326334

but it is believed to aid in enzyme binding to insoluble

substrates (Demidyuk et al., 2008).
A partial ORF, found downstream of mvp1, had highest
sequence similarity with the MprII aminopeptidase from P.
piscicida. In P. piscicida the mprII gene is found 137-nt
downstream of the previously mentioned vibriolysin encoding gene mprI (Miyamoto et al., 2002). Another partial ORF,
found upstream of mvp1, was identical with a sequence of a
predicted proline aminopeptidase from Moritella sp. PE36,
which is not found in association with a predicted vibriolysin.
The results of the present study indicate that MvP1 is a
vibriolysin, which causes necrotic and hemorrhagic tissue
damage and degrades host components. It is therefore
concluded that MvP1 may have an important role in M.
viscosa invasiveness and in establishing infection.
Acknowledgments
We wish to thank Islandsbleikja (Samherji hf) for
supplying salmon, Bergljot Magnadottir for isolated IgM
and IgM antibodies, and Unnur Steingrimsdottir for
lumpsh collagen. The EPC and BF-2 cells were a generous
gift from the Danish Institute for Food and Veterinary
Research. We also thank Jon Oskar Jonsson for his help
with temperature effect assays. This work was supported
by the AVS research fund, the Eimskip fund of the
University of Iceland and the Rannis research fund.
References
Benediktsdottir, E., Helgason, S., Sigurjonsdottir, H., 1998. Vibrio spp.
isolated from salmonids with shallow skin lesions and reared at
low temperature. J. Fish Dis. 21, 1928.
Benediktsdottir, E., Verdonck, L., Sproer, C., Helgason, S., Swings, J., 2000.
Characterization of Vibrio viscosus and Vibrio wodanis isolated at
different geographical locations: a proposal for reclassication of
Vibrio viscosus as Moritella viscosa comb. nov. Int. J. Syst. Evol. Microbiol. 50, 479488.
Booth, B.A., Boesman-Finkelstein, M., Finkelstein, R.A., 1984. Vibrio cholerae hemagglutinin/protease nicks cholera enterotoxin. Infect.
Immun. 45, 558560.
Bruno, D.W., Grifths, J., Petrie, J., Hastings, T.S., 1998. Vibrio viscosus in
farmed Atlantic salmon Salmo salar in Scotland: eld and experimental observations. Dis. Aquat. Organ. 34, 161166.
Chang, A.K., Park, J.W., Lee, E.H., Lee, J.S., 2007. The N-terminal propeptide
of Vibrio vulnicus extracellular metalloprotease is both an inhibitor
of and a substrate for the enzyme. J. Bacteriol. 189, 68326838.
David, V.A., Deutch, A.H., Sloma, A., Pawlyk, D., Ally, A., Durham, D.R.,
1992. Cloning, sequencing and expression of the gene encoding the
extracellular neutral protease, vibriolysin, of Vibrio proteolyticus. Gene
112, 107112.
Demidyuk, I.V., Gasanov, E.V., Sana, D.R., Kostrov, S.V., 2008. Structural
organization of precursors of thermolysin-like proteinases. Protein J.
27, 343354.
Denkin, S.M., Nelson, D.R., 2004. Regulation of Vibrio anguillarum empA
metalloprotease expression and its role in virulence. Appl. Environ.
Microbiol. 70, 41934204.
Finkelstein, R.A., Boesman-Finkelstein, M., Chang, Y., Hase, C.C., 1992.
Vibrio cholerae hemagglutinin/protease, colonial variation, virulence,
and detachment. Infect. Immun. 60, 472478.

Gudmundsdottir, B.K., 1996. Comparison of extracellular proteases produced by Aeromonas salmonicida strains, isolated from various sh
species. J. Appl. Bacteriol. 80, 105113.
Gudmundsdottir, B.K., Bjornsdottir, B., 2007. Vaccination against atypical
furunculosis and winter ulcer disease of sh. Vaccine 25, 55125523.
Gudmundsdottir, B.K., Hvanndal, I., Bjornsdottir, B., Wagner, U., 2003.
Analysis of exotoxins produced by atypical isolates of the sh pathogen Aeromonas salmonicida, by enzymatic and serological methods. J.
Fish Dis. 26, 1529.
Heidarsdottir, K.J., Gravningen, K., Benediktsdottir, E., 2008. Antigen
proles of the sh pathogen Moritella viscosa and protection in sh.
J. Appl. Microbiol. 104, 944951.
Komori, Y., Nonogaki, T., Nikai, T., 2001. Hemorrhagic activity and muscle
damaging effect of Pseudomonas aeruginosa metalloproteinase (elastase). Toxicon 39, 13271332.
Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly
of the head of bacteriophage T4. Nature 227, 680685.
Lee, S.O., Kato, J., Nakashima, K., Kuroda, A., Ikeda, T., Takiguchi, N.,
Ohtake, H., 2002. Cloning and characterization of extracellular metal
protease gene of the algicidal marine bacterium Pseudoalteromonas
sp. strain A28. Biosci. Biotechnol. Biochem. 66, 13661369.
Liu, P.V., 1957. Survey of haemolysin production among species of Pseudomonas. J. Bacteriol. 74, 718727.
Lunder, T., Evensen, O., Holstad, G., Hastein, T., 1995. Winter ulcer in the
Atlantic salmon Salmo salar. Pathological and bacteriological investigations and transmission experiments. Dis. Aquat. Organ. 23, 3949.
Magnadottir, B., 1998. Comparison of immunoglobulin (IgM) from four
sh species. Icel. Agric. Sci. 12, 5163.
Marchler-Bauer, A., Anderson, J.B., Derbyshire, M.K., DeWeese-Scott, C.,
Gonzales, N.R., Gwadz, M., Hao, L., He, S., Hurwitz, D.I., Jackson, J.D.,
Ke, Z., Krylov, D., Lanczycki, C.J., Liebert, C.A., Liu, C., Lu, F., Lu, S.,
Marchler, G.H., Mullokandov, M., Song, J.S., Thanki, N., Yamashita,
R.A., Yin, J.J., Zhang, D., Bryant, S.H., 2007. CDD: a conserved domain
database for interactive domain family analysis. Nucleic Acids Res. 35,
D237240.
Miyamoto, K., Tsujibo, H., Nukui, E., Itoh, H., Kaidzu, Y., Inamori, Y., 2002.
Isolation and characterization of the genes encoding two metalloproteases (MprI and MprII) from a marine bacterium, Alteromonas sp.
strain O-7. Biosci. Biotechnol. Biochem. 66, 416421.
Miyoshi, S., 2006. Vibrio vulnicus infection and metalloprotease. J. Dermatol. 33, 589595.
Miyoshi, S., Shinoda, S., 2000. Microbial metalloproteases and pathogenesis. Microbes Infect. 2, 9198.
Miyoshi, S., Wakae, H., Tomochika, K., Shinoda, S., 1997. Functional
domains of a zinc metalloprotease from Vibrio vulnicus. J. Bacteriol.
179, 76067609.
Miyoshi, S., Sonoda, Y., Wakiyama, H., Rahman, M., Tomochika, K., Shinoda, S., Yamamoto, S., Tobe, K., 2002. An exocellular thermolysin-like
metalloprotease produced by Vibrio uvialis: purication, characterization, and gene cloning. Microb. Pathog. 33, 127134.
Rawlings, N.D., Barrett, A.J., 1995. Evolutionary families of metallopeptidases. Methods Enzymol. 248, 183228.
Rawlings, N.D., Morton, F.R., Kok, C.Y., Kong, J., Barrett, A.J., 2008. MEROPS:
the peptidase database. Nucleic Acids Res. 36, D320325.
Silva, A.J., Pham, K., Benitez, J.A., 2003. Haemagglutinin/protease expression and mucin gel penetration in El Tor biotype Vibrio cholerae.
Microbiology (Reading, England) 149, 18831891.
Tang, B., Nirasawa, S., Kitaoka, M., Marie-Claire, C., Hayashi, K., 2003.
General function of N-terminal propeptide on assisting protein folding and inhibiting catalytic activity based on observations with a
chimeric thermolysin-like protease. Biochem. Biophys. Res. Commun.
301, 10931098.
Wu, Z., Milton, D., Nybom, P., Sjo, A., Magnusson, K.E., 1996. Vibrio cholerae
hemagglutinin/protease
(HA/protease)
causes
morphological
changes in cultured epithelial cells and perturbs their paracellular
barrier function. Microb. Pathog. 21, 111123.
Yeats, C., Rawlings, N.D., Bateman, A., 2004. The PepSY domain: a regulator of peptidase activity in the microbial environment? Trends
Biochem. Sci. 29, 169172.